With the development of computer science, more and more people are accustomed to effect the expression of text messages by a computer. Widely used inputting text means nowadays includes the keyboard, the touchscreen, the writing pad and so on, those means of inputting text have their respective disadvantages. Accordingly, designing a more naturalized, convenient and effective way of writing text in the air is important as a field of research.
However, as for a computer vision, the trace of a text writing in the air is consecutive, although the conventional text recognition technology is readily capable of recognizing the text consecutively written, the precision of recognition is yet to be improved.